178 ALASKAN GLACIER STUDIES similar way, it is nevertheless a special case and its advance under snowfall variation in due to peculiar individual characteristics. So far as known the Yakutat Bay region is the only one in the world where a number of glaciers of different forms, lengths and reservoir conditions have undergone a spasmodic and well graded advance and complete transformation. To account for this by climatic variation calls for a far greater increase in snowfall than has hitherto been recorded in glacier regions. Being unprecedented does not make it impossible, to be sure, but it adds one more difficulty in the way of acceptance of the hypothesis. We find it difficult also to harmonize the climatic hypothesis with the fact that the phenomena of spasmodic advance are, so far as we know, localized in and near the Yakutat Bay region. So great advance as is observed here would demand notable increase in snowfall for a period of certainly a few years; and we are unable to understand why there should be this great increase in one area, and nothing corresponding to it in the adjacent regions, where mountains rise equally high, where snowfields are equally extensive, and where glaciers are equally large. One difficulty in understanding this leads to skepticism when we remember that the region hi question is the one which was most severely shaken by the earthquakes of September, 1899. In fact, in view of all the considerations discussed above, we feel that the climatic hypothesis is improbable, that it is inadequate to explain all the facts, and that it is such a weak rival hypothesis to that of earthquake shaking that it cannot be entertained. Hypothesis of Response to Ablation. Wastage of the lower part of a glacier, as well as addition to the reservoirs must have an effect on glacier variation. It is possible, for example, that the removal of load from a lower part of a glacier might induce a responsive flowage from the heavily loaded part of the glacier above snow line of a nature similar to the flowage resulting from overloading in the reservoir. It is certain that unloading has been in progress in the lower portions of all the advancing glaciers of the Yakutat Bay region. If the advance were confined to a single glacier, or to a group of glaciers of essentially the same characteristics, this hypothesis might be applicable. But it does not seem applicable to a case in which so many glaciers of such different characteristics are involved. For instance, that underloading in the moraine-covered bulb of Atrevida Glacier should result in an advance hi the same years as the advance of Haenke Glacier whose end is in a mountainous valley seems improbable. It is equally difficult to understand by this theory why the Marvine Glacier, which terminates in the piedmont Malaspina Glacier, advanced in the same year as the Atrevida; and also why the other tributaries to the Malaspina have not responded to unloading of the Malaspina by ablation. Thus, even without considering whether the cause itself is really an adequate possibility for such a spasmodic advance in a single glacier, and without making use of the fact that no such response to unloading is known ever to have occurred, the hypothesis fails and may be discarded hi so far as its application to the advancing glaciers of the Yakutat Bay region is concerned. Hypothesis of Elevation and Tilting. Increase in snow supply to glaciers could be brought about by increase in elevation, and this theory has also been considered. Here again we cannot apply the test of actual measurement, since only the approximate elevations of a few of the peaks are known; but this theory, like that of increase in snowfall, lacks probability. Applying it, for example, to Marvine or Lucia Glaciers, whose sources.